International Journal of Heat and Mass Transfer 54 (2011) 5527–5539

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

journal homepage: www.elsevier.com/locate/ijhmt

Heat transfer from a hot moving cylinder impinged by a planar subcooled water jet
M. Gradeck a,⇑, A. Kouachi a, J.L. Borean b, P. Gardin b, M. Lebouché a
a
LEMTA, Nancy University – CNRS, 2 avenue de la forêt de haye 54504, Vandoeuvre lès Nancy, France
b
AM Research, voie romaine 57283, Maizières lès Metz, France

a r t i c l e i n f o a b s t r a c t

Article history: A hot moving (rotating) cylinder was heated up to 500–600 °C and then was cooled by a planar water jet
Received 10 December 2010 impinging on a line parallel to the symmetry axis. The time dependent wall temperature was measured
Received in revised form 11 July 2011 using embedded thermocouples and the corresponding wall heat fluxes were estimated through an
Accepted 20 July 2011
inverse conduction method. In a recent paper, we showed that cooling rates depend on the subcooled
Available online 16 August 2011
temperature of the jet, the velocity of the jet and the surface-to-jet velocity ratio. Since the initial tem-
perature of the cylinder was higher than the Leidenfrost temperature, we observed all the boiling regimes
Keywords:
from film boiling to nucleate boiling. The objectives of this paper are firstly to describe the current con-
Heat transfer
Inverse solution
ditions which exist in the Run Out Table in hot rolling mills, secondly to review the main experimental
Heat conduction studies dedicated to jet cooling which have led to modelling heat transfer in boiling conditions and finally
Impinging jet to propose new correlations taking into account the velocity of the wall.
Transient boiling Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction be between 200 and 500 °C, depending on steel grade). At the very
beginning of the cooling phase, the temperature of the steel strip is
In the steel industry, cooling on the Run Out Table (ROT) after above the Leidenfrost temperature so film boiling, transition boil-
hot rolling is one of the most difficult process steps in the hot mill ing, critical heat flux (CHF) and nucleate boiling all occur. Control-
strip. The decrease in the temperature needs to be perfectly con- ling the cooling rate and the homogeneity thereof thus remains a
trolled because the mechanical properties of steel alloys are condi- major challenge for manufacturers aiming to produce steels with
tioned by the cooling rate ensured by these jets [1–4]. Generally, desired and homogeneous mechanical properties.
top cooling is carried out using a number of subcooled water jets
which impinge perpendicularly on the hot steel surface while bot-
1.1. Description of the flow on the Run Out Table (ROT)
tom cooling is done using sprays. The water jets are organized in a
set called a header where two jets rows are either aligned or stag-
As previously described, the heat transfer can not be homoge-
gered. Complex flows are thus obtained, as a result of the interac-
neous when using water jets in the cooling system because of flow
tion between the jets and the moving surface. In-depth knowledge
topology and because the water film depth above the hot surface is
on the heat transfer associated to that flow is therefore essential
not constant. This depends on the distance from the impact zone of
including knowledge of the interaction between the jet and the
the jet but also on the ratio between velocity of the jet and the
moving surface, the interaction between the jets and the interac-
velocity of the moving surface. In a recent paper, Gradeck et al.
tion between the ramps.
[5] carried out experimental and numerical studies of the flow
Rates of cooling will vary between 15 and 1000 K/s depending
structure of a single water jet impinging on a moving surface. This
on the required steel mechanical property. Although cooling tech-
work provided a valuable correlation to predict the position of the
nologies based on the so-called laminar water jets have been
hydraulic water jump for operating conditions similar to those in
widely studied in the past, knowledge of these technologies re-
ROT cooling systems. In a more recent paper, the flow pattern of
mains incomplete which means it is difficult to attain optimum
multiple water jet impinging on a moving surface was numerically
production. Obviously the kinetics of cooling depends on the vari-
studied by Cho et al. [6] using the CFD Fluent package. Their com-
ous boiling regimes met during transient cooling (i.e. the metal
putations clearly showed how flow patterns are dependant on the
slab is reheated before rolling, the temperature of the strip after
running conditions (flow rates, velocity of the surface). At low flow
rolling is about 900 °C and after cooling, the temperature should
rates, hydraulic jumps were observed while with increasing values
of the flow rates, the hydraulic jumps disappeared and a pool was
⇑ Corresponding author. observed. Moreover, a fountain effect was found to occur at times
E-mail address: michel.gradeck@ensem.inpl-nancy.fr (M. Gradeck). between two adjacent jets with consequent improvement of the

0017-9310/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijheatmasstransfer.2011.07.038
5528 M. Gradeck et al. / International Journal of Heat and Mass Transfer 54 (2011) 5527–5539

Nomenclature

a heat transfer coefficient (W m2 K1) r⁄ dimensionless velocity, VVSJ

d nozzle diameter (m) x distance from the jet axis (m)
q00 heat flux (W m2) x⁄ dimensionless distance, dx
Re Reynolds number
Pr Prandtl number Subscripts
DTsub subcooled temperature (K) FB film boiling
DTsat superheat temperature (K) MFB minimum of film boiling
T temperature (°C) TB transition boiling
VJ jet velocity at the impingement (m s1) NB nucleate boiling
Vn jet velocity at the exit of the nozzle (m s1) TC transient conduction
Vs wall velocity (m s1) CHF critical heat flux
k conductivity (W m1 K1) L liquid
l dynamic viscosity (Pa s) V vapour
r surface tension (N m1)

cooling flux. Monde et al. [7] gave an illustration of the interactions
of neighbouring jets and the expected increase of the heat flux in
this area. However in more recent study, Franco [8] did not find
a fountain effect to occur whether the alignment of the jets was
staggered or aligned. The large distance between jets (around
100 mm) in Franco’s [8] experiments is the probable explanation.
1.2. Heat transfer associated with single impinging jet
Most studies of heat transfer associated with impinging jets
have been carried out using a static surface which means that
the dynamic conditions as mentioned above may have been far
from industrial conditions [9–15]. However, most correlations
from these studies are used as command and control for the ROT
even though the influence of surface velocity was not addressed
in these studies. Depending on the strip temperature, four different
water cooling regimes may be encountered: (i) for high tempera-
tures, only film boiling; (ii) transition boiling if the surface temper-
ature of the strip is lower than the Leidenfrost temperature; (iii)
nucleate boiling regime if the surface temperature is lower than
the CHF temperature and finally; (iv) forced convection
(Fig. 1(a)). A particular shape of boiling curve may be observed in
the transition regime in the case of an impinging jet on a hot plate
(Fig. 1(b)). After CHF, a first minimum of flux was observed after
which the heat flux increased again and reached a high value. This
‘‘shoulder of flux’’ was found to falls down abruptly to film boiling
regime for very high superheats. Miyasaka et al. [15], Ishigai et al.
[9], Ochi et al. [10] and Hall [11], Robidou et al. [13] and Gradeck et
al. [16] have reported the existence of a ‘‘shoulder of flux’’ beneath
the jet (transition boiling regime) in the case of static surface. But
Gradeck et al. [16] showed that heat transfer is radically modified
when the impingement surface is moving. They found that the
‘‘shoulder of flux’’ beneath the jet axis collapsed and that the local
boiling curves had the same shape upstream, downstream and be-
neath the centreline of the jet.
(or models) available for impinging jets in these four boiling
regimes.
1.3.1. Film boiling
1.3.1.1. Stagnation zone. Zumbrunnen et al. [20,21] and Filipovic
[22] obtained models for the film boiling regime by solving the
boundary layer equations but as some strong assumptions have
been assumed to simplify the Navier–Stokes equations and obtain
an analytical solution, these models can be very far from the cur-
rent application. For example neglecting waviness on the va-
pour–liquid interface as well as Kelvin–Helmotz instabilities lead
to delay the minimum of film boiling (in comparison with
measurements).
Liu and Wang [23,24] derived an analytical expression of the
heat flux for the film boiling regime in the stagnation zone
(C = 1.414). To fit the experiments, the constant C of the original
expression was slightly modified to take into account the interfa-
cial waviness of the vapor layer (C = 2):
.
Q 00FB ¼ CRe0;5 0166
J Pr J ðkL kV DT sub DT sat Þ0;5 d; ð1Þ
where the thermal properties of water were evaluated at the film
temperature of water and vapor properties evaluated at the film
temperature of vapour.
Ochi et al. [10] proposed the following correlation
q00FB ¼ 3:18x105 ð1 þ 0:383DT sub ÞðV J =dÞ0:828 ð2Þ
A similar expression was found by Ishigai et al. [9] for a planar jet:
q00FB ¼ 5:4x104 ð1 þ 0:527DT sub ÞV 0:607
J : ð3Þ
1
For all of these correlations V is expressed in m s , d is in mm and
DT in K.
Robidou [25] used an innovative experimental device to mea-
sure boiling curves under steady state conditions and proposed
the following correlation at the stagnation zone:

1.3. ROT heat transfer model q00FB ¼ 5:38x104 ð5:5 þ DT sub ÞV 0:6 ð4Þ
J

Given that cooling rates (i.e., heat transfer at the wall) are af-
fected by the temperature of the strip, dedicated correlation needs
be applied to the off-line or on-line models used in ROT in order to
correctly predict the coiling temperature of the strip. Since the heat
transfer model used in the ROT control is usually simple, consisting
of a simple polynomial relationship [17–19], improvements should
be done to consider four heat transfer regimes (Fig. 1) and two sig-
nificant temperatures (Leidenfrost and CHF temperatures). In the
following sub-sections, we give an overview of the correlations
1.3.1.2. Parallel flow zone. Few models exist in the parallel flow
zone (x⁄ > 2). Filipovic [22] developed a model for this regime
based on the boundary layers equations and measurements
achieved on a parallel jet. Hatta et al. [26] proposed the following
correlation fitting the experimental data of Kokado et al. [27] ob-
tained for a transient cooling:
a ¼ 200ð2420  21:7T L ÞDT 0:8
sat : ð5Þ
 M. Gradeck et al. / International Journal of Heat and Mass Transfer 54 (2011) 5527–5539 5529

(a)

Forced Nucleate Transition Filmboiling

convection boiling boiling

Heat flux (W/m²)

Critical heatflux

Leiden frost point

ΔTsat = Tp-Tsat (K) (K)

(b)

Forced Nucleate Transition Film boiling

convection boiling boiling
Heat flux (W/m²)

Critical heat flux

Leiden frost point

ΔTsat = Tp-Tsat (K)

Fig. 1. Shape of the boiling curve in impinging jet experiments: (a) parallel flow zone of the jet, (b) impinging zone of the jet.

This correlation underestimated the heat transfer coefficient in T MFB ¼ a  bT L ; ð8aÞ

comparison with the values measured by Robidou [25] at the min-
imum of film boiling (MFB). with a = 1100 °C, b = 8.5 for Kokado’s experiment.
The previous relationship must be used according to the critical
1.3.2. Temperature for the minimum film boiling (MFB) temperature, Tc (which was 68 °C in the case of Kokado’s experi-
Several results can be found concerning the temperature for the ments): (i) If TL < Tc, a wetting regime occurs; (ii) if TL > Tc and
MFB (i.e. TMFB) in the stagnation zone. Liu [24] gave an expression TW < TMFB, a partial wetting regime occurs; (iii) if TL > Tc and
for the temperature of minimum film boiling, TMFB. TW > TMFB, film boiling can occur. Following on from these results,
  under the jet and for water temperature below 40 °C, the wall is al-
DT MFB ¼ 14DT sub Pr0;33
J ðkL =kV Þ lL =lV ð6Þ ways in a wetted regime.
Hatta and Osakabe [28] used the water temperature whereas
And the associated heat flux:
Robidou [25] used the subcooling:
 0;5
q00FB;h ¼ 0:5DT MFB kV Re0;5 0166
J Pr J lL = lV =d; ð7Þ
T MFB ¼ 326 þ 17:6DT 0:8
sub ð8bÞ
where the thermal properties of water were evaluated at the film
temperature of water and vapor properties evaluated at the film A change of TMFB according to the hydrodynamics of the jet-wall
temperature of vapor. Following Liu [24], TMFB depends only on interaction needs to be taken into account in this expression.
water properties and subcooling and hydrodynamics does not have In a recent paper, Karwa et al. [29] have also developed a sim-
any effect (see expression 6), Kokado et al. [27]: plified two-phase analytical model for predicting of the TMFB in
5530 M. Gradeck et al. / International Journal of Heat and Mass Transfer 54 (2011) 5527–5539

the case of a planar or a circular liquid jet impinging on a flat hot Miyasaka et al. [15] found the following relationship to estimate
wall. They provided the following equation: CHF for a planar jet impinging a static plate:

DT MFB ¼ 0:7DT sub Pr0;5 0;5

L ReL ðkL =kV ÞðdV =wÞ for a planar jet; ð9aÞ q0CHF ¼ 0; 16hlv qv ð1 þ 0; 86V 0;38
n Þ
 0;8  1;13 ! 
0;5
ql cpl DT sub rgðql  qv Þ 0;25
DT MFB ¼ 1:6DT sub Pr 0;5
L ðkL =kV ÞðlL =lV Þ for a circular jet; ð9bÞ  1 þ 0112
qv h lv q2v
:

where the properties of vapour are taken at the bulk temperature, w ð12bÞ
is the width of the jet just before impingement and dV, the vapour
thickness. Although the hydrodynamic parameters used are not too far from
the requirements of the ROT, several points prevent the extension
1.3.3. Transition boiling of these correlations to the ROT including:
Transition boiling is an important regime because it is at the
origin of control issues. When the strip temperature decreases, ? The surface is fixed.
the extracted heat flux increases and the temperature discrepan- ? Experiments are dependent on the size of heater, which is a
cies are amplified because of high discrepancies on local heat flux. parameter of the correlations.
A definition of the transition boiling was provided by Berenson
[30], Kalinin et al. [31] and Pan et al. [32] namely a combination 1.3.5. Nucleate boiling
of unstable film boiling and unstable nucleate boiling alternately Wolf [12], Wolf et al. [37] proposed correlation (9) for nucleate
existing at any given location on a heating surface. The extracted boiling:
heat flux in the transition boiling was provided by Kalinin et al.
q00NB ¼ 63:7DT 2:95
sat : ð13Þ
[31]:
q00TC tTC þ q00NB t NB þ q00FB t FB Following Wolf’s finding [12], the previous correlation can also be
q00TB ¼ ð10Þ applied in a parallel flow configuration:
tTC þ t NB þ tFB
With TB, transition boiling; TC, transient conduction; NB, saturated q0NB ¼ aDT bsat : ð14Þ
nucleate boiling; FB, film boiling;t, fraction of time, contact time of This correlation means that heat transfer does not depend on
each regime. hydrodynamics in this regime.
For Nagai and Nishio [33], Ohtake and Koizumi [34], transition Filipovic et al. [38,39] proposed the following correlation (for a
boiling is a stationary boiling regime and the heat flux is expressed static surface) taking into account the velocity of the jet, subcool-
as: ing, the diameter of the nozzle and superheat. However this corre-
q00TB ¼ Cq00NB þ ð1  CÞq00FB ð11Þ lation also overestimates the heat flux in the case of nucleate
boiling for moving system:
In fact both approaches lead to the last equation with the weighting
0:608
parameter having a different meaning when transient conduction is q0NB ¼ C ðDT sub ÞV 0:5
j d DT 0:14
sat ; ð15Þ
assumed as negligible. In the case of time-dependant models, the 7
where C(DTsub = 80 °C) = 1.3  10 and C(DTsub = 65 °C) = 0.81 
weighting C is based on contact time whereas it is based on the li-
107; Vj in m/s, d in mm.
quid–solid contact fraction area in stationary models. These ap-
proaches are very interesting because expressions can be adapted
1.4. Concluding remarks on our literature review
to the case of the ROT using experiments. The current parameters
used are obtained for nuclear reactor applications by using droplets,
The real process is extremely complicated because of the multi-
pool boiling experiments, vertical tubes, saturated fluid, tempera-
ple interactions involved namely the interaction of the jet with the
tures below 300 °C, with surfaces (copper, sapphire, etc.) whose
band, of a jet with another jet or of a header with another header.
thermal properties are very different from those of steel. These can-
This means it is impossible to cover all these situations at the same
not be used for the Run-Out Table without prior adaptations. More-
time in the context of a simple laboratory experiment. Hence pre-
over the moving effect of the surface is not taken into account at all
vious studies have always been carried out using simple configura-
although Gradeck et al. [16] found this significantly modifies the
tions such as multiple jet impinging a cold plate, a single jet
boiling curve and transition boiling regime.
impinging a moving cold plate or a single jet impinging one plate
whose temperature was higher than the Leidenfrost temperature.
1.3.4. Critical heat flux, CHF
Previous theoretical approaches also used a simple situation as
Monde et al. [7] and his team (Mitsutake et al. [35], Hammad
their starting point, for example an axisymmetric jet impinging a
et al. [36]) have carried out a lot of experiments in the field of
static cold surface or planar jet impinging a moving plate (cold
impinging jets and they have provided a correlation to estimate
or hot). For our study, we opted to experiment with a new situa-
the CHF:
8 pffiffiffiffiffiffiffiffiffiffiffi tion, namely a planar jet impinging a hot moving surface (i.e. a
>
> q00
¼
1þ 1þ4CJa rotating cylinder). The Table 1 gives an overview of the running
> qCHF
> 2
;
>
> 0
conditions for the different experimental works which have been
>
> q CpDT
< Ja ¼ qVL hLVsub ; previously mentioned.
 0:645  0:343 ; ð12aÞ
>
> q0 ¼ 0:221 qqL 2r
ð1 þ D=dÞ0:364 ;
>
> qL V J ðDdÞ
>
>
V 2. Experimental set-up and data reduction
>
> 0:95ðd=DÞð1þD=dÞ0:364
:C ¼
ðq =q Þ0:43 ð2r=q V ðDdÞÞ0:343
L V L J
The experimental device was designed to allow local and instan-
where D was the size of the heater and thermal properties were taneous determination of the heat transfer coefficient between a
evaluated at the saturated temperature. This correlation is applica- water jet impinging a moving surface. This estimation was quite
ble for stationary applications, static surface and to a qqL of 5.3–1603, accurate because we used temperature measurements of the wall
V
the inverse Weber number (second term in brackets) of 2  107 to surface and calculating the cooling heat flux by means of an inverse
3
10 and D/d of 5–30. conduction method [40]. This device is illustrated in Fig. 2.
 M. Gradeck et al. / International Journal of Heat and Mass Transfer 54 (2011) 5527–5539 5531

Table 1
Experimental conditions of the cited works.

Authors Experiment Liquid Jet Surface Running conditions Regimes

2
Ishigai et al. [9] Transient Water Free (20 or 80)  12 mm 1 6 VJ 6 3.2 m/s Nucleate boiling
Planar Stainless steel 5 6 DTsub 6 55 K Transition boiling
50  6.2 mm2 Film boiling
Ochi et al. [10] Transient Water Free 210  50 mm2 2 6 VJ 6 7 m/s Nucleate boiling
Circular Stainless steel 5 6 DTsubB 6 80 K Transition boiling
5, 10 and 20 mm Film boiling
Kokado et al. [27] Transient Water Free 2000  10 mm2 1 6 QV 6 7 l/min Wetting
Circular Stainless steel TL = 20 °C Non wetting
Ø10 mm
Mitsutake et al. [35] Transient Water Free Width 4 mm 5 6 VJ 6 60 m/s CHF
Circular Thickness 0.1 mm 0.1 6 P 6 1.3 MPa
Ø2 mm Stainless steel TL = 20 °C
Ni
Wolf [12] Stationary Water Free 119  35.7 mm2 10 6 x 6 90 mm Forced convection
Imposed heat flux Planar Alloy Ni–Cr–W–Mo 2 6 VJ 6 5 m/s Onset of nucleate boiling
10  105 mm2 DTsub = 50 K
0.25 6 q00 6 6.34 MW/m2 Nucleate boiling
Miyasaka et al. [15] Stationary Water Free 4  8 mm2 1.1 6 VJ 6 15.3 m/s Forced convection
Planar Pt TL = 15 °C Nucleate boiling
10  32 mm2 CHF
Miyasaka et al. [15] Stationary Water Free Ø1.5 mm 1.1 6 VJ 6 15.3 m/s Nucleate boiling
Planar Pt TL = 15 °C CHF
10  32 mm2 Transition boiling
Robidou et al. [13] Stationary Water Free 10  104 mm2 0.46 6 Vn 6 0.9 m/s All regimes
Controlled temperature Planar Ni on Cu 5 6 DTsub 6 17 K
1  9 mm2
Kouachi [41], Transient Water Free Rotating cylinder 0.32 6 Vn 6 0.69 m/s All regimes
Gradeck et al. [16]
Planar Ni 0.5 6 r⁄ 6 1.25 m/s
4  180 mm2 Ø175 mm 18 6 DTsub 6 50 K

Inspection 5 Planar Jet

Glass
6 6
8 8
ω

7
24T 9 10

11

T: temperature measurement
Q: flow rate measurement
ω: velocity measurement (rpm)
P: power measurement

T
Q

Fig. 2b. Details of the experimental set-up [41].

2 3
1
Fig. 2a. Schematic of the experimental set-up [41].
2.1. Experimental set-up
The cooling of the steel strip in the ROT can be approximated
using a rotating hot cylinder which respects the main process
parameters of the cooling to some extent namely the speed of
the moving surface, its temperature and the jet characteristics
4
(flow rate and water temperature). The results obtained in this
study of heat transfer beneath the jet and close to the impact zone
(in these zones, the effect of the curvature is negligible) can used
for ROT characterisation. However, this device cannot be used to
simulate the effect of the accumulation of water because in the real
process a certain amount of water can be transported by the steel
strip. Of course, water accumulation is also a parameter that
should be considered in future works because heat transfer may
be different for a free jet or a submerged jet.
5532 M. Gradeck et al. / International Journal of Heat and Mass Transfer 54 (2011) 5527–5539
Sensor signal
T 2Τ
Temperature 2π or x*=0 2π
Heat flux
Fig. 3. Building of the boiling curves.
The set-up is schematically outlined in Fig. 2a and photos
show details of the device in Fig. 2b. The boiling vessel contained
the jet (5) and the hot rotating cylinder (7). The external
diameter of the outer cylinder (which is cooled by the jet) was
equal to 175 mm and its internal diameter equals 100 mm. The
length of this cylinder was 200 mm. Heating was provided by
three electrical wires which were wound on an inner cylinder
(Øout ¼ 100 mm;Øin ¼ 85 mm) which itself was adjusted with
the outer cylinder. Heating with a constant heat flux (P) was thus
applied on the internal radius until thermal equilibrium was
reached (uniform wall temperature required for the test of about
600 °C on the external radius). Each resistance used can dissipate
2000 W (10). The wall temperature was almost uniform because
heat losses (radiation and convection) were almost uniform (i.e.
the cylinder rotates during the heating phase). Two cooled stain-
less steel flasks (6) were set out on both sides to prevent the re-
volving contacts, ball bearings and sealing rings from excessively
high temperatures. This was also used to center the cylinder and
thus avoid the unbalances. The system was allowed to dilate
without generating additional constraints inside the set-up. Axial
losses (i.e. axial conduction) are impossible to avoid but as the
cylinder was long enough and thermocouples placed in the mid-
dle of it, we only considered a 2D conduction problem to solve
the inverse problem related to the estimation of heat flux [40].
24 thermocouples (N type) were inserted near the external sur-
face in grooves and a coating was deposited. The electrical output
was provided to resistances by means of revolving contacts (8a).
The test fluid (water) was first heated at a subcooled temperature
in a vessel (1) and circulated in a primary circuit. The flow rate
was adjusted using a pump (2) and controlled using an electro-
magnetic flowmeter (3). When all the initial parameters were ad-
justed, the electromagnetic sluice gate (4) was opened and the
cooling of the cylinder could start. The jet had an outlet cross
sectional area of 180  4 mm2 and the water jet outlet was lo-
cated at a distance H from the hot cylinder and impinged the
higher line of the cylinder. This distance could be adjusted. The
pre-amplified signals of the thermocouples as well as their angu-
lar positions (i.e. a position sensor was available on the drive-
shaft) were then also transmitted to a computer (11) via
revolving contacts (8b) to be recorded. This meant we could con-
struct the time evolution of temperature at each point on the
surface of the cylinder. More details on this device can be found
in [41].
3T time =
q’’ (W/m²)
2π position
time
ΔTSAT
time
2.2. Data reduction
When the cylinder reached the equilibrium temperature, the
heating was stopped and the jet cooling began at t = 0 (opening
of the sluicegate). During cooling, the heat flux was estimated
through inversion of the temperature measurements using an In-
verse Heat Conduction Model (IHCM); see Volle et al. [40]. This
model was based on an analytical solution of the 2D transient heat
equation. As axial conduction was neglected, we therefore worked
with a 2D problem. The heat equation can be solved using a cylin-
drical reference with knowledge of the initial temperature field
(i.e. an equilibrium temperature field obtained with a constant
heat flux P at an internal radius with losses due to convection
and radiation at the external surface and conduction at each end
of cylinder). The use of spatial integral transform, see Maillet
et al. [42], namely the Fourier transform (noted), led us to obtain
a linear relation (in the Fourier space) between the kth harmonic of
inside temperature1 and the kth harmonic of heat flux lost on the
cooled wall:
~ kq
T~ k ¼ X ~k ; ð16Þ
~ k , the sensitivity matrix linking variations of temperature re-
with X
sponses to variations of wall heat flux input. Once T~ k was estimated
through experimental measurements, q ~k could be calculated thanks
to an inversion of the previous equation. Because of the standard
noise deviation of experimental temperatures, an ordinary least
square operator was used to obtain the most accurate estimation
of the cooling heat flux in Fourier domain:
 1
~k ¼ X
q ~
~T X ~ T T~ k :
X ð17Þ
k k k
A Fourier inversion of the previous estimated signal led to the cool-
ing flux identification as a function of time. Obviously the inverse
problem is an ill-posed problem and it was thus necessary to use
some regularization techniques. In our study, the probes were lo-
cated close to the wall (500 lm depth) and we used a Beck future
time regularization method to obtain a better estimation. The accu-
racy of the method was estimated using noisy data (i.e. tempera-
ture) which was obtained either using the analytical solution of
the heat equation or numerical simulations. Any errors would
1
The thermocouples were inserted at 500 lm depth.
 M. Gradeck et al. / International Journal of Heat and Mass Transfer 54 (2011) 5527–5539 5533

depend on the choice of hyperparameters or regularization param- heat flux (Fig. 4c), first minimum of film boiling and shoulder of
eters (number of future times, number of harmonics) and of course flux (Fig. 4a).
on the noise level [43,44].
As we could know the position of the thermocouple as a func-
tion of time, it was then possible to link the time evolution of 3. Heat transfer analysis and correlations
the temperature at a given point to the time-evolution of heat flux
at the same point and thus to build the boiling curve corresponding 3.1. Heat transfer analysis
to a curvilinear abscissa x⁄ (Fig. 3). An example of the data obtained
(local boiling curves) is given in the Fig. 4. Thus we were able to As described in a previous paper [16], heat transfer does not
identify the different heat transfer regimes namely forced convec- evolve in the same way as in a static case and its value is clearly
tion, onset of nucleate boiling, nucleate boiling (Fig. 4b), critical modified when the impingement surface is moving. As a

6
x 10
3
x* = 6.9
x* = 9.15
2.5 x* = 11.45
x* = 13.75
x* = 16
2
Heat flux (W/m²)

1.5

0.5

0
0 50 100 150 200 250 300 350 400 450
DTsat (K)
(a)
5 6
x 10 x 10
3.5
x* = 6.9 x* = 6.9
16
x* = 9.15 CHF at MHF x* = 9.15
x* = 11.45 x* = 11.45
14
Heat flux (W/m²)

3
Heat flux (W/m²)

x* = 13.75 x* = 13.75
nucleate boiling x* = 16 x* = 16
12

10 ONB 2.5

8
2
6 Forced convection

4
1.5
2

0 1
10 20 30 40 50 60 70 40 50 60 70 80 90 100 110 120 130
DTsat (K) DTsat (K)

(b) (c)

legend
Fig. 4. Local boiling curves from [16], DTsub = 34 K; uj = 1,06 m/s  u⁄ = 1,25.
5534 M. Gradeck et al. / International Journal of Heat and Mass Transfer 54 (2011) 5527–5539

because the jet was spreading over a stable vapour film. Neverthe-
less, in a recent published paper [16], we especially shown a col-
q’’ (W/m²)

Increasing the surface velocity

First minimum
Fig. 5. Shape of the local boiling curves with a moving surface.
consequence, the maximum of CHF is lower than in the static case
[25]. But the most obvious effect was a kind of auto-similarity of
the local boiling curves throughout the cooling zone. In fact, defin-
ing a local boiling curve at x⁄ for a moving surface integrates all the
boiling phenomena over the surface throughout the cooling zone
upstream. Bubbles or vapour spot moved with the surface (for
nucleate boiling and transition boiling regimes) and in the film
boiling regime, the stability of the vapour layer is clearly hard to
achieve which is why a kind of ‘‘shoulder of flux’’ was observed.
For a static surface, the shoulder of flux was observed at the
impingement of the jet (x⁄ = 0) but collapsed in the flow zone
lapse of the ‘‘shoulder of flux’’ beneath the jet axis and a
displacement of the maximum of heat transfer downstream. An-
other point is that the local boiling curves have slightly the same
shape upstream, downstream and beneath the point of maximum
heat transfer (i.e. maximum heat transfer was not always at the
centreline of the jet; see Gradeck et al. [16]). The shape of the boil-
ing curve looked rather like those beneath the impingement
(Fig. 5). Seiler-Marie et al. [45] have modelled the transition, from
transition boiling to the shoulder of flux and this is based on the
assumption of the existence of periodic bubble oscillations con-
Leidenfrost temperature
trolled by Rayleigh–Taylor instabilities namely that vapour spots
would be fragmented by the jet hydrodynamic forces. In fact, for
ΔTSAT (K) moving surfaces, rewetting the wall and thus stability of the va-
pour spots or film should be also controlled by Kelvin–Helmholtz
instabilities.
Another point to be recalled is related to the position of the
cooling zone (i.e. the zone corresponding to a temperature de-
crease): at the beginning of cooling, the cooling zone begins mainly
beneath the jet and develops only downstream [16]. As a conse-
quence, the maximum of heat transfer was not located beneath
the jet and it moves during the cooling. An asymmetry of the heat
transfer was also observed upstream and downstream from the
maximum of heat transfer (Fig. 4). This was a consequence of the
jet-wall interaction, especially when a counter current appeared
(upstream) and induced rapid evolution of the velocity flow field.
The data bank obtained by Kouachi [41] was analyzed and heat
transfer correlations for nucleate boiling and maximum heat

4 4
Vj = 1,06 m/s; r* = 0,5; ΔTsub = 18 K x=0 Vj = 1,06 m/s; r* = 1; DTsub = 34 K
x= 4,5 Vj = 1,06 m/s; r* = 1; DTsub = 18 K
3.5 3.5
x = 9,15

3 3
Heat flux (MW/m²)
Heat flux (MW/m2)

2.5 2.5

2 2

1.5 1.5

1 1

0.5 0.5

0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
ΔTsat (K) ΔΤsat (Κ)

(a) local boiling curves (b) Influence of subcooling, x/d = 9.15

4 4
Vj = 1,20m/s; Vs = 1,14 m/s; DTsub = 50 K
Vj = 1,06 m/s; ΔTsub = 50 K
r* = 0,5
3.5 3.5
Vj = 1,04m/s; Vs = 1,14 m/s; DTsub = 50 K r* = 1
r* = 1,25
3 3
Heat flux (Mw/m²)

Heat flux (MW/m²)

2.5 2.5

2 2

1.5 1.5

1 1

0.5 0.5

0 0
0 20 40 60 80 100 0 20 40 60 80 100
ΔTsat (K) ΔTsat (K)

(c) Influence of jet velocity, x/d = 9.15 (d) Influence of speed surface, x/d = 8
Fig. 6. Local evolution of boiling curves in nucleate boiling regime.
 M. Gradeck et al. / International Journal of Heat and Mass Transfer 54 (2011) 5527–5539 5535

2.5

+20%

Calculated heat flux (MW/m²)

2 -20%

1.5

0.5

0
0 0.5 1 1.5 2 2.5
Experimental data (MW/m²)

Fig. 7. Comparison between experimental data and correlation (18).

6
4
Vj = 1,06 m/s; DTsub = 50 K r* = 0,5 Vj = 1,06 m/s; r* = 1,25
r* = 0,75 DTsub = 50 K
r* = 1 3.5 DTsub = 34 K
5 r* = 1,25
DTsub =18 K
3
4
CHF (MW/m²)

2.5
CHF (MW/m²)

3 2

1.5
2

Vs 1
1
0.5

0 0
0 2 4 6 8 10 12 14 16 18 0 2 4 6 8 10 12 14 16 18
x/d x/d

(a) influence of the surface velocity (b) influence of the subcooling

6 6
Vs = 1,14 m/s; DTsub = 50 K Vj = 0,69 m/s; DTsub = 50 K
H = 50 mm
Vj = 1,04 m/s
5 H = 10 mm
5 Vj = 1,20 m/s

4 4
CHF (MW/m²)

CHF (MW/m²)

3 3

2 2

1 1

0 0
0 2 4 6 8 10 12 14 16 18 0 2 4 6 8 10 12 14 16 18
x/d x/d

(c) influence of the jet velocity (d) influence of the nozzle distance
Fig. 8. Local evolutions of the CHF.

transfer were proposed taking into account the main hydrody- 3.2. Heat transfer associated with moving surface
namic parameters, the fluid properties, the fluid subcooling and
also the dissymmetry of the heat transfer between upstream flow 3.2.1. Nucleate boiling regime
and downstream flow. Further experiments would be required to The heat flux at the onset of boiling was independent of
study other regimes like transition boiling or film boiling. curvilinear position but slightly increased with subcooling. In the
5536 M. Gradeck et al. / International Journal of Heat and Mass Transfer 54 (2011) 5527–5539

15

14
+12%
13

12

calculated x/d
11

10 -12%

5
5 7 9 11 13 15
experimental x/d

Fig. 9. Comparison between data and calculated xCHF;max using (21).

6
+20%
5
calculated CHF (MW/m²)

4 -20%

0
0 1 2 3 4 5 6
experimental CHF (MW/m²)

Fig. 10a. Comparison between data and calculated CHF using (22), upstream.

+20%
5
calculated CHF (MW/m²)

4 -20%

0
0 1 2 3 4 5 6
experimental CHF (MW/m²)

Fig. 10b. Comparison between data and calculated CHF using (22), downstream.
 M. Gradeck et al. / International Journal of Heat and Mass Transfer 54 (2011) 5527–5539 5537

12
Wolf [38]
Gradeck et al.
10

Heat flux (MW/m²)

8

0
20 30 40 50 60 70
ΔTsat (K)

Fig. 11a. Comparison of the correlations (12) and (17) for nucleate boiling.

25
Miyasaka et al. [15]
Gradeck et al.
20
Heat flux (MW/m²)

15

10

0
0 20 40 60 80 100
ΔTsub (K)

Fig. 11b. Comparison of the correlations (12b) and (22) for CHF, x⁄ = 0, r⁄ = 0.

nucleate boiling regime, the curvilinear position and subcooling
was found to have little influence on the cooling flux (Fig. 6a and
b).
Fig. 6c and d, respectively show the influence of jet velocity and
the influence of wall velocity on the boiling curves. The cooling flux
at the onset of boiling varied slightly with the wall velocity but in-
creased quicker with the speed of the jet. The cooling flux in the
fully developed nucleate boiling regime seemed independent of
jet velocity and wall velocity. However, it should be noted that
near the CHF (DTsat around 80–100 K), the dependence of the cool-
ing flux on the speed surface became significant, Fig. 6d.
This means that in the fully developed nucleate boiling regime
heat transfer was found not to depend on hydrodynamics which
was expected. This means our data could be fitted with a power
law of DTsat. The correlation that best fits our data in the nucleate
boiling regime is:
q0FNB ¼ 2; 46DT 1;64
sat
where q0FNB (kW/m2) and 30 < DTsat < 60 K.
A comparison of the data is shown in the Fig. 7. The closer we
got to the CHF, the less data fitted with the Eq. (17).
3.2.2. CHF
The CHF can be observed only for x⁄ > 0 [16]. No cooling was
measured for negative values of x⁄. Fig. 8a and b show the
evolution of local CHF for various conditions of speed and
subcooling.
The CHF was not found to be uniform throughout the cooling
zone and thus there was a maximum that decreased when the
wall speed was increased (Fig. 8a). Moreover, the position of that
maximum increased when the wall speed was increased and
moved towards high x⁄ values. Another point to be noted was
that the CHF was also controlled by the water subcooling. In
Fig. 8b, we plotted its spatial evolution for three values of
subcooling.
Fig. 8c and d show the evolution of CHF for different jet veloci-
ties (i.e., increasing the distance between nozzle and wall was
equivalent to increase the jet velocity).
We can define two kinds of correlations as follows:
ð18Þ
 
xCHF;max ¼ f1 V j ; V s ; DT sub ð19Þ
 
q0CHF ¼ f2 V j ; V s ; DT sub ; x 
ð20Þ
5538 M. Gradeck et al. / International Journal of Heat and Mass Transfer 54 (2011) 5527–5539

A dimensional analysis gave the following relationship: up and used to obtain the time-dependent surface temperature
 0794  0209  0247 !0134 and extracted heat flux. The findings of this experimental study
xCHF;max Vs H qv hlv could be used for control cooling systems in ROT. The main find-
¼ 1; 0015We0202 Ja0579 ;
d Vn d ql V 2n ings of this study are as follows:
ð21Þ
c DT (1) The cooling rates were significantly different for moving sur-
with We ¼ 0036  0; 17; Ja ¼ pl h sub ¼ 0033  0088; V s =V n ¼ 1; 36  3;
lv faces as compared with static surface. We observed the same
56; H=d ¼ 2; 5  12; 5; qv =ql ¼ 8; 4  105  33  105 ; hlv =V 2n ¼ 5
6 6 spatial pattern with film boiling, shoulder of flux, transition
10  23  10 .
boiling, CHF and finally nucleate boiling.
Fig. 9 compares the positions of the maximum CHF determined
(2) The main difference was observed in the shoulder of flux
experimentally with those calculated using the previous Eq. (19).
region where the moving of the surface induced a collapse
This shows that the equation correlates the data with a maximum
of the heat flux.
error of ±12%.
(3) Correlations for nucleate boiling and CHF are available for
As we knew the position of the maximum CHF, we could esti-
wide range of operating conditions. Surface velocity is a very
mate the value of CHF upstream and downstream of this position.
sensitive parameter for CHF.
The correlation given by Miyasaka et al. [15] could be corrected by
(4) Finally, we conclude that further experiments are needed to
a factor taking into account the effect of the speed surface (r⁄) as
explain the collapse of the shoulder of flux and to predict the
well as the decrease of CHF from its maximum value:
  rewetting (TMFB) of the surface for high superheats.
x < xCHF;max : q0CHF ¼ 0; 186hlv qv 1 þ 0; 86V 0;38
n
 0;56 !  
q rg ql  qv 0;25 References
 1 þ 0; 328 l Ja
qv 2 qv
 [1] S. Serajzadeh, Prediction of temperature distribution and phase transformation
x 0317 on the run-out table in the process of hot strip rolling, Appl. Math. Model. 27
 1þ ð1 þ r Þ1926 ð22aÞ (2003) 861–875.
d
[2] Y.K. Sun, D. Wu, Effect of ultra-fast cooling on microstructure of large section
  bars of bearing steel, J. Iron. Steel Res. Int. 16 (5) (2009) 61–65. 80.
x > xCHF;max : q0CHF ¼ 0; 186hlv qv 1 þ 0; 86V 0;38
n [3] R.W.K. Honeycombe, Steels: Microstructure and Properties, Edward Arnold,
  !  0;25 London, 1981.
0;56
ql rg ql  qv [4] N. Hatta, J. Kokado, H. Takuda, J. Harada, K. Hiraku, Modelling for cooling
 1 þ 0; 328 Ja process of a hot plate by a laminar water bar, Steel Res. 55 (1984) 143–148.
qv q2v [5] M. Gradeck, A. Kouachi, A. Dani, D. Arnoult, J.L. Borean, Experimental and
 x 0;0486 numerical study of the hydraulic jump of an impinging jet on a moving surface,
 1þ ð1 þ r  Þ0;83 ð22bÞ Exp. Therm. Fluid Sci. 30 (2006) 193–201.
d [6] M.J. Cho, B.G. Thomas, P.J. Lee, Three-dimensional numerical study of
⁄
with Vn = 0,32  0,69 m/s; DTsub = 18  50 K et r = 0,5  1,25. impinging water jets in runout table cooling processes, Metall. Mater. Trans.
B 39 (2008) 4.
The correlations correspond to our experimental data with a [7] M. Monde, H. Kusuda, H. Uehara, Burnout heat flux in saturated forced
maximum deviation of ±20% (Figs. 10a and 10b). convection boiling with two or more impinging jets, Trans. JSME 46 (1980)
1834–1843.
[8] G. Franco, Boiling heat transfer during cooling of a hot moving steel plate by
3.3. Comparisons with existing correlations multiple top jets, Ph.D., University of British Columbia, Vancouver, December
2008.
It is clear that heat fluxes were strongly modified in all the boil- [9] S. Ishigai, A. Nakanishi, T. Ochi, Boiling heat transfer for a plane water jet
impinging on a hot surface, in: 6th International Heat Transfer Conference, vol.
ing regimes when the wall was moving. If we compare our results 1, FB 30, 1978, pp. 445–450.
with the correlations given in Section 1, the correlations used so far [10] T. Ochi, S. Nakanishi, M. Kaji, S. Ishigai, Multi-phase and Heat Transfer III. Part
for static surfaces lead to an overestimation of heat transfers for a A: Fundamentals – Cooling of a Hot Plate with an Impinging Circular Water Jet,
Elsevier Science Publishers B.V., Amsterdam, 1984. pp. 671–681.
moving surface. Figs. 11a and 11b show our comparisons of the [11] E.D. Hall, F.P. Incropera, R. Viskanta, Jet impingement boiling from a circular
proposed correlations for nucleate boiling (Eq. (18)) and CHF (Eq free-surface jet during quenching: part 1. Single-phase jet, J. Heat Transfer 123
22) with the findings of Wolf et al. [37] (Eq. (13)) and Miyasaka (2001) 901–910.
[12] D.H. Wolf, Turbulent development in a free surface jet and impingement
et al. [15] (Eq. (12b)). The question was why heat transfer is over-
boiling heat transfer, Ph.D., Purdue University, 1993.
estimated by previous correlations? It is surprising that heat trans- [13] H. Robidou, H. Auracher, P. Gardin, M. Lebouché, Controlled cooling of a hot
fer is widely modified for nucleate boiling because as we showed in plate with a water jet, Exp. Therm. Fluid Sci. 26 (2-4) (2002) 123–129.
[14] Aloke Kumar Mozumder, Masanori Monde, Peter Lloyd Woodfield, Md.
Section 1, most of the correlations used for nucleate boiling do not
Ashraful Islam, Maximum heat flux in relation to quenching of a high
usually take any hydrodynamics effects into account. To our temperature surface with liquid jet impingement, Int. J. Heat Mass Transfer 49
knowledge, only the equation proposed by Filipovic et al. [38,39] (17–18) (2006) 2877–2888.
(correlation for a static surface) takes into account the velocity of [15] Y. Miyasaka, S. Inada, Y. Owase, Critical heat flux and subcooled nucleate
boiling in transient region between a two-dimensional water jet and a heated
the jet, the subcooling, the diameter of the nozzle and superheat surface, J. Chem. Eng. Jpn. 13 (1980) 29–35.
but this also leads to the heat flux in the case of nucleate boiling [16] M. Gradeck, A. Kouachi, M. Lebouché, F. Volle, D. Maillet, J.L. Borean, Boiling
for moving system being overestimated. curves in relation to quenching of a high temperature moving surface with
liquid jet impingement, Int. J. Heat Mass Transfer 52 (5–6) (2009) 1094–1104.
CHF was also overestimated by the correlation given by Mya- [17] Yi. Zheng, Li. Shaoyuan, Xiaobo Wang, Distributed model predictive control for
saka et al. [15]. In that particular case, this is less surprising be- plant-wide hot-rolled strip laminar cooling process, J. Process. Control 19
cause as Critical Heat Flux occurs, the amount of vapour which is (2009) 1427–1437.
[18] Ananya Mukhopadhyay, Sudipta Sikdar, Implementation of an on-line run-out
produced at the wall may be carried by the moving wall which table model in a hot strip mill, J. Mater. Process. Technol. 169 (2005) 164–172.
leads to heat transfer being limited. [19] H.B. Xie, Z.Y. Jiang, X.H. Liu, G.D. Wang, A.K. Tieu, Prediction of coiling
temperature on run-out table of hot strip mill using data mining, J. Mater.
Process. Technol. 177 (2006) 121–125.
4. Concluding remarks [20] D.A. Zumbrunnen, F.P. Incropera, R. Viskanta, Method and apparatus for
measuring heat transfer distributions on moving and stationary plates cooled
This paper presents new results regarding the water cooling of a by a planar liquid jet, Exp. Therm. Fluid Sci. 3 (1990) 202–213.
[21] D.A. Zumbrunnen, F.P. Incropera, R. Viskanta, The effect of surface motion on
hot moving strip which was initially at a temperature around forced convection film boiling heat, J. Heat Transfer ASME 111 (1989) 760–
600 °C. Special instrumentations and signal treatment were set 766.
 M. Gradeck et al. / International Journal of Heat and Mass Transfer 54 (2011) 5527–5539
[22] J.V. Filipovic, Film boiling over a moving horizontal surface, Ph.D. Purdue
University, 1994.
[23] Z.-H. Liu, J. Wang, Study on film boiling heat transfer for water jet impinging
on high temperature flat plate, Int. J. Heat Mass Transfer 44 (2001) 2475–
2481.
[24] Z-H. Liu, Prediction of minimum heat flux for water jet boiling on a hot plate, J.
Thermophys. Heat Transfer 17 (2003) 2.
[25] H. Robidou, Etude expérimentale du refroidissement diphasique à haute
température par jet d’eau impactant, Thèse de doctorat, Université Henri
Poincaré, Nancy I, 2000.
[26] N. Hatta, J-I. Kokado, H. Takuda, J. Harada, K. Hiraku, Predictable modeling for
coolings process of a hot steel plate by a laminar water bar, Arch.
Eisenhüttenwes. 55 (1984). Nr.4.
[27] J.-I. Kokado, N. Hatta, H. Takuda, J. Harada, N. Yasuhira, An analysis of film
boiling phenomena of subcooled water spreading radially on hot steel plate,
Arch. Eisenhüttenwes. 55 (3) (1984). März.
[28] N. Hatta, H. Osakabe, Numerical modeling for cooling process of a moving hot
plate by a laminar water curtain, ISIJ Int. 29 (11) (1995) 919–925.
[29] N. Karwa, T. Gambaryan-Roisman, P. Stephan, C. Tropea, A hydrodynamic
model for subcooled liquid jet impingement at the Leidenfrost condition, Int. J.
Therm. Sci. (2011), doi:10.1016/j.ijthermalsci.2011.01.021.
[30] P.J. Berenson, Film boiling heat transfer from a horizontal surface, J. Heat
Transfer 83 (3) (1961) 351–358.
[31] E.K. Kalinin, I.I. Berlin, V.V. Kostiouk, Transition boiling heat transfer, Advances
in Heat Transfer, vol. 18, Academic Press, 1987.
[32] Pan Chin, J.Y. Hwang, T.L. Lin, The mechanism of heat transfer in transition
boiling, Int. J. Heat Mass Transfer 32 (7) (1989) 1337–1349.
[33] N. Nagai, S. Nishio, Leidenfrost temperature on an extremely smooth surface,
Exp. Therm. Fluid Sci. 12 (1996) 373–379.
[34] H. Ohtake, Y. Koizumi, Derivations of correlation and liquid–solid contact
model of transition boiling heat transfer, JSME Int. J. Ser. B (2) (2006).
5539
[35] Y. Mitsutake, M. Monde, Ultra high critical heat flux during forced flow boiling
heat transfer with an impinging jet, J. Heat Transfer 125 (6) (2003) 1038–1045.
[36] J.A. Hammad, Y. Mitsutake, M. Monde, Movement of maximum heat flux and
wetting front during quenching of hot cylindrical block, Int. J. Therm. Sci. 43
(8) (2004) 743–752.
[37] D.H. Wolf, F.P. Incropera, R. Viskanta, Local jet impingement boiling heat
transfer, Int. J. Heat Transfer 39 (7) (1996) 1395–1406.
[38] J. Filipovic, R. Viskanta, F.P. Incropera, Cooling of a moving steel strip by an
array of round jets, in: 35th MWSP Conference Proceedings, ISS-AIME, vol.
XXXI, 1994, p. 317.
[39] J. Filipovic, R. Viskanta, F.P. Incropera, T.A. Veslocki, Cooling of a moving steel
strip by an array of round jets, Steel Res. (12) (1994).
[40] F. Volle, D. Maillet, M. Gradeck, A. Kouachi, M. Lebouché, Practical application
of inverse heat conduction for wall condition estimation on a rotating cylinder,
Int. J. Heat Mass Transfer 52 (1-2) (2009) 210–221. 15.
[41] A. Kouachi, Étude expérimentale de l’ébullition convective d’un jet d’eau plan
impactant une surface mobile portée à hautes températures, Thèse de
Doctorat de l’Université de Nancy 1, Nancy, 2006. ddc: 620. <http://
www.scd.uhp-nancy.fr/docnum/SCD_T_2006_0136>.
[42] D. Maillet, S. André, J.C. Batsale, A. Degiovanni, C. Moyne, Thermal
Quadrupoles: Solving the Heat Equation through Integral Transforms, John
Wiley and Sons, UK, 2000.
[43] F. Volle, D. Maillet, M. Gradeck, M. Lebouché, Semi-analytical inverse heat
conduction on a rotating cylinder with Laplace and Fourier transforms, Inverse
Problem in Science and Engineering-IPSE-Cambridge 16 (5) (2008) 655–674.
[44] F. Volle, M. Gradeck, A. Kouachi, D. Maillet, M. Lebouché, Inverse heat
conduction applied to the measurement of heat fluxes on a rotating cylinder:
comparison between an analytical and a numerical technique, J. Heat Transfer
130 (8) (2008), doi:10.1115/1.2928013.51-52.081302.
[45] N. Seiler-Marie, J.M. Seiler, O. Simonin, Transition boiling at jet impingement,
Int. J. Heat Mass Transfer 47 (2004) 5059–5070.